How stereolabs/zed-sdk works
Project Type and Purpose
• This repository is an open-source sample collection for the ZED SDK, demonstrating how to use Stereolabs’ ZED stereo cameras in C++, Python, C#, and C.
• It is organized by feature modules (Depth Sensing, Body Tracking, Object Detection, etc.) plus a set of tutorials, and contains sample applications and exporters (JSON, FBX, KML, SVO).
File-Structure Analysis
• Top-Level Folders:
• tutorials
• camera control
• camera streaming
• depth sensing
• positional tracking
• global localization
• spatial mapping
• object detection
• body tracking
• recording
• zed one
• Each feature folder contains language-specific subfolders (cpp, python, csharp, sometimes c) with build scripts (CMakeLists.txt), source (src/), headers (include/), and README.
• Special Submodules:
– global localization/map server (a small web server serving map tiles)
– global localization/live, playback, recording (fusion with GNSS)
– exporters (JSON, FBX, KML, SVO)
– camera streaming sender/receiver (network sockets)
Insights from README
• The README defines the SDK’s feature modules, supported languages, and integration points.
• Emphasizes real-time performance, multi-language API, and external integrations (OpenGL, OpenCV, ROS, Unity).
Essential Details to Capture
• The multi-language nature (C++, Python, C#, C).
• The modularity: one core SDK with many independent sample demos.
• Data flows: raw camera frames → processing → visualization / export / network.
• Build artifacts: CMake / Visual Studio / pip.
• Networking lanes for streaming and web map server.